For the sensor data, there are numerous way to explore and build models. In this chapter, we mainly focus on classification; however, there are several aspects which would need deeper exploration, especially the following:
- Training data represents a time-ordered flow of events, but we are not going to reflect the time information but look at the data as one complete piece of information
- The same for test data -a single activity event is a part of an event stream captured during performing an activity and it can be easier to categorize it with knowledge of actual context
Nevertheless, for now, we ignore the time dimension and apply classification to explore possible patterns in the sensor data which would characterize performed activities.